-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCUDAlatticeBoltzmann2D.cu
352 lines (308 loc) · 13.3 KB
/
CUDAlatticeBoltzmann2D.cu
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
// #include <pycuda-complex.hpp>
// #include <surface_functions.h>
#include <stdint.h>
#include <cuda.h>
#define I2D(ni,i,j) (((ni)*(j)) + i)
#define T0_f(t_j,nj) (tempMax + t_j*(tempMin-tempMax)/(nj-1))
// #define pi 3.14159265f
// textures on device //
texture<float, cudaTextureType2D, cudaReadModeElementType> tex_f1, tex_f2, tex_f3, tex_f4, tex_f5, tex_f6, tex_f7, tex_f8;
texture<float, cudaTextureType2D, cudaReadModeElementType> tex_g1, tex_g2, tex_g3, tex_g4, tex_g5, tex_g6, tex_g7, tex_g8;
////////////////////////////////////////////////////////////////////////////////
////////////////////// STREAM //////////////////////////
////////////////////////////////////////////////////////////////////////////////
__global__ void stream_kernel ( float *f1_in, float *f2_in, float *f3_in, float *f4_in, float *f5_in, float *f6_in, float *f7_in, float *f8_in,
float *g1_in, float *g2_in, float *g3_in, float *g4_in, float *g5_in, float *g6_in, float *g7_in, float *g8_in){
int t_i = blockIdx.x*blockDim.x + threadIdx.x;
int t_j = blockIdx.y*blockDim.y + threadIdx.y;
int tid = t_i + t_j*blockDim.x*gridDim.x;
// look up the adjacent f's needed for streaming using textures
// i.e. gather from textures, write to device memory: f1_in, etc
f1_in[tid] = tex2D(tex_f1, (float) (t_i-1) , (float) t_j);
f2_in[tid] = tex2D(tex_f2, (float) t_i , (float) (t_j-1));
f3_in[tid] = tex2D(tex_f3, (float) (t_i+1) , (float) t_j);
f4_in[tid] = tex2D(tex_f4, (float) t_i , (float) (t_j+1));
f5_in[tid] = tex2D(tex_f5, (float) (t_i-1) , (float) (t_j-1));
f6_in[tid] = tex2D(tex_f6, (float) (t_i+1) , (float) (t_j-1));
f7_in[tid] = tex2D(tex_f7, (float) (t_i+1) , (float) (t_j+1));
f8_in[tid] = tex2D(tex_f8, (float) (t_i-1) , (float) (t_j+1));
g1_in[tid] = tex2D(tex_g1, (float) (t_i-1) , (float) t_j);
g2_in[tid] = tex2D(tex_g2, (float) t_i , (float) (t_j-1));
g3_in[tid] = tex2D(tex_g3, (float) (t_i+1) , (float) t_j);
g4_in[tid] = tex2D(tex_g4, (float) t_i , (float) (t_j+1));
g5_in[tid] = tex2D(tex_g5, (float) (t_i-1) , (float) (t_j-1));
g6_in[tid] = tex2D(tex_g6, (float) (t_i+1) , (float) (t_j-1));
g7_in[tid] = tex2D(tex_g7, (float) (t_i+1) , (float) (t_j+1));
g8_in[tid] = tex2D(tex_g8, (float) (t_i-1) , (float) (t_j+1));
}
////////////////////////////////////////////////////////////////////////////////
////////////////////// COLLIDE //////////////////////////
////////////////////////////////////////////////////////////////////////////////
__global__ void collide_kernel (int ni, int nj, float beta, float tau, float tauT, float faceq1, float faceq2, float faceq3,
float *f0_in, float *f1_in, float *f2_in, float *f3_in, float *f4_in, float *f5_in, float *f6_in, float *f7_in, float *f8_in,
float *g0_in, float *g1_in, float *g2_in, float *g3_in, float *g4_in, float *g5_in, float *g6_in, float *g7_in, float *g8_in,
float g, float tempMin, float tempMax, int plotTemp, float *plot_data){
int t_i = blockIdx.x*blockDim.x + threadIdx.x;
int t_j = blockIdx.y*blockDim.y + threadIdx.y;
int tid = t_i + t_j*blockDim.x*gridDim.x;
float rtau = 1.f/tau;
float rtau1 = 1.f - rtau;
float rtauT = 1.f/tauT;
float rtau1T = 1.f - rtauT;
// Read all f's and g's and store in registers
float f0now = f0_in[tid];
float f1now = f1_in[tid];
float f2now = f2_in[tid];
float f3now = f3_in[tid];
float f4now = f4_in[tid];
float f5now = f5_in[tid];
float f6now = f6_in[tid];
float f7now = f7_in[tid];
float f8now = f8_in[tid];
float g0now = g0_in[tid];
float g1now = g1_in[tid];
float g2now = g2_in[tid];
float g3now = g3_in[tid];
float g4now = g4_in[tid];
float g5now = g5_in[tid];
float g6now = g6_in[tid];
float g7now = g7_in[tid];
float g8now = g8_in[tid];
// Macroscopic flow props:
float ro = f0now + f1now + f2now + f3now + f4now + f5now + f6now + f7now + f8now;
float T = g0now + g1now + g2now + g3now + g4now + g5now + g6now + g7now + g8now;
float vx = (f1now - f3now + f5now - f6now - f7now + f8now)/ro;
float vy = (f2now - f4now + f5now + f6now - f7now - f8now)/ro;
// Set plotting variable to velocity magnitude
if (plotTemp == 1) plot_data[tid] = T;
/* else plot_data[tid] = 0.3f*sqrtf(vx*vx + vy*vy);*/
else plot_data[tid] = 0.2f*vy;
// plot_data[tid] = T;
// Calculate equilibrium f's
float v_sq_term = 1.5f*(vx*vx + vy*vy);
float f0eq = ro * faceq1 * (1.f - v_sq_term);
float f1eq = ro * faceq2 * (1.f + 3.f*vx + 4.5f*vx*vx - v_sq_term);
float f2eq = ro * faceq2 * (1.f + 3.f*vy + 4.5f*vy*vy - v_sq_term);
float f3eq = ro * faceq2 * (1.f - 3.f*vx + 4.5f*vx*vx - v_sq_term);
float f4eq = ro * faceq2 * (1.f - 3.f*vy + 4.5f*vy*vy - v_sq_term);
float f5eq = ro * faceq3 * (1.f + 3.f*(vx + vy) + 4.5f*(vx + vy)*(vx + vy) - v_sq_term);
float f6eq = ro * faceq3 * (1.f + 3.f*(-vx + vy) + 4.5f*(-vx + vy)*(-vx + vy) - v_sq_term);
float f7eq = ro * faceq3 * (1.f + 3.f*(-vx - vy) + 4.5f*(-vx - vy)*(-vx - vy) - v_sq_term);
float f8eq = ro * faceq3 * (1.f + 3.f*(vx - vy) + 4.5f*(vx - vy)*(vx - vy) - v_sq_term);
// Calculate equilibrium g's
// float g0eq = T * faceq1 * (1.f );
// float g1eq = T * faceq2 * (1.f + 3.f*vx );
// float g2eq = T * faceq2 * (1.f + 3.f*vy );
// float g3eq = T * faceq2 * (1.f - 3.f*vx );
// float g4eq = T * faceq2 * (1.f - 3.f*vy );
// float g5eq = T * faceq3 * (1.f + 3.f*(vx + vy) );
// float g6eq = T * faceq3 * (1.f + 3.f*(-vx + vy) );
// float g7eq = T * faceq3 * (1.f + 3.f*(-vx - vy) );
// float g8eq = T * faceq3 * (1.f + 3.f*(vx - vy) );
float g0eq = T * faceq1 * (1.f - v_sq_term);
float g1eq = T * faceq2 * (1.f + 3.f*vx + 4.5f*vx*vx - v_sq_term);
float g2eq = T * faceq2 * (1.f + 3.f*vy + 4.5f*vy*vy - v_sq_term);
float g3eq = T * faceq2 * (1.f - 3.f*vx + 4.5f*vx*vx - v_sq_term);
float g4eq = T * faceq2 * (1.f - 3.f*vy + 4.5f*vy*vy - v_sq_term);
float g5eq = T * faceq3 * (1.f + 3.f*(vx + vy) + 4.5f*(vx + vy)*(vx + vy) - v_sq_term);
float g6eq = T * faceq3 * (1.f + 3.f*(-vx + vy) + 4.5f*(-vx + vy)*(-vx + vy) - v_sq_term);
float g7eq = T * faceq3 * (1.f + 3.f*(-vx - vy) + 4.5f*(-vx - vy)*(-vx - vy) - v_sq_term);
float g8eq = T * faceq3 * (1.f + 3.f*(vx - vy) + 4.5f*(vx - vy)*(vx - vy) - v_sq_term);
// Do collisions (no temperature effect)
// f0_in[tid] = rtau1 * f0now + rtau * f0eq;
// f1_in[tid] = rtau1 * f1now + rtau * f1eq;
// f2_in[tid] = rtau1 * f2now + rtau * f2eq;
// f3_in[tid] = rtau1 * f3now + rtau * f3eq;
// f4_in[tid] = rtau1 * f4now + rtau * f4eq;
// f5_in[tid] = rtau1 * f5now + rtau * f5eq;
// f6_in[tid] = rtau1 * f6now + rtau * f6eq;
// f7_in[tid] = rtau1 * f7now + rtau * f7eq;
// f8_in[tid] = rtau1 * f8now + rtau * f8eq;
float T0 = T0_f(t_j,nj);
//Apply temperature gradient effect
f0_in[tid] = rtau1 * f0now + rtau * f0eq;
f1_in[tid] = rtau1 * f1now + rtau * f1eq;
f2_in[tid] = rtau1 * f2now + rtau * f2eq - 3*beta*faceq2*(T-T0)*g;
f3_in[tid] = rtau1 * f3now + rtau * f3eq;
f4_in[tid] = rtau1 * f4now + rtau * f4eq + 3*beta*faceq2*(T-T0)*g;
f5_in[tid] = rtau1 * f5now + rtau * f5eq - 3*beta*faceq3*(T-T0)*g;
f6_in[tid] = rtau1 * f6now + rtau * f6eq - 3*beta*faceq3*(T-T0)*g;
f7_in[tid] = rtau1 * f7now + rtau * f7eq + 3*beta*faceq3*(T-T0)*g;
f8_in[tid] = rtau1 * f8now + rtau * f8eq + 3*beta*faceq3*(T-T0)*g;
g0_in[tid] = rtau1T * g0now + rtauT * g0eq;
g1_in[tid] = rtau1T * g1now + rtauT * g1eq;
g2_in[tid] = rtau1T * g2now + rtauT * g2eq;
g3_in[tid] = rtau1T * g3now + rtauT * g3eq;
g4_in[tid] = rtau1T * g4now + rtauT * g4eq;
g5_in[tid] = rtau1T * g5now + rtauT * g5eq;
g6_in[tid] = rtau1T * g6now + rtauT * g6eq;
g7_in[tid] = rtau1T * g7now + rtauT * g7eq;
g8_in[tid] = rtau1T * g8now + rtauT * g8eq;
}
////////////////////////////////////////////////////////////////////////////////
////////////////////// APPLY BC //////////////////////////
////////////////////////////////////////////////////////////////////////////////
__global__ void apply_BCs_kernel (int ni, int nj, float vxin, float roout, float faceq1, float faceq2, float faceq3,
float *f0_in, float *f1_in, float *f2_in, float *f3_in, float *f4_in, float *f5_in, float *f6_in, float *f7_in, float *f8_in,
float *g0_in, float *g1_in, float *g2_in, float *g3_in, float *g4_in, float *g5_in, float *g6_in, float *g7_in, float *g8_in,
float tempMin, float tempMax, int* solid_data){
// Apply all BC's apart from periodic boundaries:
int t_i = blockIdx.x*blockDim.x + threadIdx.x;
int t_j = blockIdx.y*blockDim.y + threadIdx.y;
int tid = t_i + t_j*blockDim.x*gridDim.x;
// Solid BC: "bounce-back"
if (solid_data[tid] == 0) {
float f1old = f1_in[tid];
float f2old = f2_in[tid];
float f3old = f3_in[tid];
float f4old = f4_in[tid];
float f5old = f5_in[tid];
float f6old = f6_in[tid];
float f7old = f7_in[tid];
float f8old = f8_in[tid];
f1_in[tid] = f3old;
f2_in[tid] = f4old;
f3_in[tid] = f1old;
f4_in[tid] = f2old;
f5_in[tid] = f7old;
f6_in[tid] = f8old;
f7_in[tid] = f5old;
f8_in[tid] = f6old;
}
if (t_j== 0) {
float f1old = f1_in[tid];
float f2old = f2_in[tid];
float f3old = f3_in[tid];
float f4old = f4_in[tid];
float f5old = f5_in[tid];
float f6old = f6_in[tid];
float f7old = f7_in[tid];
float f8old = f8_in[tid];
f1_in[tid] = f3old;
f2_in[tid] = f4old;
f3_in[tid] = f1old;
f4_in[tid] = f2old;
f5_in[tid] = f7old;
f6_in[tid] = f8old;
f7_in[tid] = f5old;
f8_in[tid] = f6old;
g0_in[tid] = faceq1 * tempMax;
g1_in[tid] = faceq2 * tempMax;
g2_in[tid] = faceq2 * tempMax;
g3_in[tid] = faceq2 * tempMax;
g4_in[tid] = faceq2 * tempMax;
g5_in[tid] = faceq3 * tempMax;
g6_in[tid] = faceq3 * tempMax;
g7_in[tid] = faceq3 * tempMax;
g8_in[tid] = faceq3 * tempMax;
}
if (t_j== nj-1 ) {
float f1old = f1_in[tid];
float f2old = f2_in[tid];
float f3old = f3_in[tid];
float f4old = f4_in[tid];
float f5old = f5_in[tid];
float f6old = f6_in[tid];
float f7old = f7_in[tid];
float f8old = f8_in[tid];
f1_in[tid] = f3old;
f2_in[tid] = f4old;
f3_in[tid] = f1old;
f4_in[tid] = f2old;
f5_in[tid] = f7old;
f6_in[tid] = f8old;
f7_in[tid] = f5old;
f8_in[tid] = f6old;
g0_in[tid] = faceq1 * tempMin;
g1_in[tid] = faceq2 * tempMin;
g2_in[tid] = faceq2 * tempMin;
g3_in[tid] = faceq2 * tempMin;
g4_in[tid] = faceq2 * tempMin;
g5_in[tid] = faceq3 * tempMin;
g6_in[tid] = faceq3 * tempMin;
g7_in[tid] = faceq3 * tempMin;
g8_in[tid] = faceq3 * tempMin;
}
if (t_j== 1) {
g0_in[tid] = faceq1 * tempMax;
g1_in[tid] = faceq2 * tempMax;
g2_in[tid] = faceq2 * tempMax;
g3_in[tid] = faceq2 * tempMax;
g4_in[tid] = faceq2 * tempMax;
g5_in[tid] = faceq3 * tempMax;
g6_in[tid] = faceq3 * tempMax;
g7_in[tid] = faceq3 * tempMax;
g8_in[tid] = faceq3 * tempMax;
}
// if (t_j== nj-2 ) {
// g0_in[tid] = faceq1 * tempMin;
// g1_in[tid] = faceq2 * tempMin;
// g2_in[tid] = faceq2 * tempMin;
// g3_in[tid] = faceq2 * tempMin;
// g4_in[tid] = faceq2 * tempMin;
// g5_in[tid] = faceq3 * tempMin;
// g6_in[tid] = faceq3 * tempMin;
// g7_in[tid] = faceq3 * tempMin;
// g8_in[tid] = faceq3 * tempMin;
// }
// // Inlet BC - very crude
// if (t_i == 0) {
// float v_sq_term = 1.5f*(vxin * vxin);
//
// f1_in[tid] = roout * faceq2 * (1.f + 3.f*vxin + 3.f*v_sq_term);
// f5_in[tid] = roout * faceq3 * (1.f + 3.f*vxin + 3.f*v_sq_term);
// f8_in[tid] = roout * faceq3 * (1.f + 3.f*vxin + 3.f*v_sq_term);
//
// }
//
// // Exit BC - very crude
// if (t_i == (ni-1)) {
// int tid_2 = tid - 1;
// f3_in[tid] = f3_in[tid_2];
// f6_in[tid] = f6_in[tid_2];
// f7_in[tid] = f7_in[tid_2];
//
// }
}
////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////
__global__ void apply_Periodic_BC_kernel (int ni, int nj,
float *f1_in, float *f2_in, float *f3_in, float *f4_in, float *f5_in, float *f6_in, float *f7_in, float *f8_in,
float *g1_in, float *g2_in, float *g3_in, float *g4_in, float *g5_in, float *g6_in, float *g7_in, float *g8_in){
int t_i = blockIdx.x*blockDim.x + threadIdx.x;
int t_j = blockIdx.y*blockDim.y + threadIdx.y;
int tid = t_i + t_j*blockDim.x*gridDim.x;
int tid_2;
//periodicas en tapas de los lados
if (t_i == 0 ) {
tid_2 = tid + ni -1;
f1_in[tid] = f1_in[tid_2];
f5_in[tid] = f5_in[tid_2];
f8_in[tid] = f8_in[tid_2];
g1_in[tid] = g1_in[tid_2];
g5_in[tid] = g5_in[tid_2];
g8_in[tid] = g8_in[tid_2];
}
if (t_i == (ni-1)) {
tid_2 = tid - ni +1;
f3_in[tid] = f3_in[tid_2];
f6_in[tid] = f6_in[tid_2];
f7_in[tid] = f7_in[tid_2];
g3_in[tid] = g3_in[tid_2];
g6_in[tid] = g6_in[tid_2];
g7_in[tid] = g7_in[tid_2];
}
//Periodicas en tapas de Arriba
// if (t_j == 0 ) {
// tid_2 = t_i + (nj-1)*pitch/sizeof(float);
// f2_in[tid] = f2_in[tid_2];
// f5_in[tid] = f5_in[tid_2];
// f6_in[tid] = f6_in[tid_2];
// }
// if (t_j == (nj-1)) {
// tid_2 = t_i;
// f4_in[tid] = f4_in[tid_2];
// f7_in[tid] = f7_in[tid_2];
// f8_in[tid] = f8_in[tid_2];
// }
}